Psychometric re-evaluation of the revised version of the Competitive State Anxiety Inventory-2

Abstract

The Competitive State Anxiety Inventory-2 (CSAI-2) has been the measure of choice among anxiety researchers in the sport domain since its development in 1990. A 17-item revised version, the CSAI-2R, was developed in 2003 by Cox and colleagues. The present study re-evaluated the psychometric characteristics of the CSAI-2R. Repeated measures data from 92 social-competitive tennis players at five points in time were subjected to principal axis factoring with promax rotation. Results supported the hypothesized, 3-factor measurement model. The measurement model was recovered cleanly at all time points, explaining between 58.1% and 68.8% of variance. Minimal complexity among items was evident. Cronbach alpha coefficients exceeded criterion values for all subscales at each time point (cognitive anxiety = .82 - .88, somatic anxiety = .79 - .89, self-confidence = .87 - .92). Inter-correlations among factors were in line with theoretical predictions and supported their conceptual independence. Overall, results supported the factorial validity of the revised scale in five replications. The present findings indicate that the CSAI-2R has satisfactory psychometric characteristics, unlike those reported by Lane and colleagues in 1999 for the original CSAI-2, which showed a flawed measurement model. Results suggest that researchers investigating anxiety in sport should use the CSAI-2R in preference to the original CSAI-2.